Research in the Unit of Neural Network Physiology is primarily concerned with functions of the neocortex and the basal ganglia. These brain regions are involved in higher cognitive functions e.g. executive decisions and working memory, as well as in movement control and reward mediated behavior. Our work focuses on the function of the neurotransmitters glutamate and dopamine at the network level. These neurotransmitters are involved in a variety of disease states. Dopaminergic and glutamatergic dysfunctions of the cortex are found in e.g. schizophrenia and epilepsy. A dopaminergic imbalance in the striatum, which is part of the basal ganglia, is correlated with severe movement disorders such as seen in Parkinson?s disease and Chorea Huntington. More specifically, research in our unit seeks to determine the function of microcircuits in the cortex and the striatum in order to address two related questions: (1) How does the cortex achieve and maintain activity states that allow it to encode and process information? (2) How are cortical network dynamics decoded by the striatum? We address these questions at the network level because many of the computational properties of the cortex and the striatum are predicted to emerge out of the collective action of thousands of neurons and cannot be predicted from the behavior of single neurons alone. We use a variety of experimental and computational techniques to monitor and analyze the dynamics of networks in vitro in organotypic co-cultures and acute slices. For example, we reconstruct parts of the cortex-basal ganglia systems in vitro by culturing young rat or mouse brains for up to several months on multi-electrode arrays. These neuronal co-cultures provide the most complex in vitro system that exists to date: a 6-layered cortical network that drives activity in a striatal network and also receives dopaminergic inputs from the substantia nigra. The system comprises of several hundred thousand of neurons and replicates network activity that strongly resembles that seen in vivo. Taking advantage of this approach, we are in the unique position to study single neuron electrophysiology, synaptic transmission between neurons, and neuronal populations within and across nuclei under in vivo-like conditions. This year?s research further developed the following two aspects of information processing in the cortex-basal ganglia system. (A) Dynamics in cortical networks We recently provided the first demonstration that cortical networks operate in a ?critical state?. At this stable state, the network is maximally excitable without being epileptic. Using multi-electrode arrays in combination with organotypic cultures and acute slices, we demonstrated that propagation of synchronized activity in the critical state takes on the form of ?neuronal avalanches?, which are neither wave-like, nor rhythmic, or random. These ?neuronal avalanches? are described by a power law with slope ?3/2 and a branching parameter of 1 at which they retain maximal information as they propagate through the network (Beggs and Plenz, 2003). During the last year, we demonstrated that ?neuronal avalanches? are highly diverse, yet temporally precise at the millisecond time scale and reoccur over many hours. They thus fulfill many of the requirements of a substrate for memory, and suggest that they play a central role in both information transmission and storage in cortex (Beggs and Plenz, 2004). Recent findings in our laboratory are related to the specific pharmacological profile for avalanche induction and the relevance of the precise spatiotemporal dynamics of neuronal avalanches for information transmission within cortical networks. (B) Striatal processing of cortical inputs Using calcium imaging from distal dendrites, we were the first to demonstrate that number of back propagating spikes controls dendritic calcium during ?up? states, the characteristic network state of the striatum in response to cortical inputs, (Kerr and Plenz, 2002). This last year, we demonstrated that the precise timing between ?up-state onset and delay to first action potential also determines dendritic calcium through an NMDA mediated mechanism (Kerr and Plenz, 2004). These findings pave the way for spike-time dependent plasticity rules in striatal processing of cortical inputs. We also demonstrated that GABAergic synapses between striatal neurons are important for processing of cortical inputs to the striatum. The specific local circuitries formed by these synapses can be classified into feedforward and feedback networks with unique temporal properties. (C) We also have several ongoing projects in which new technologies are combined to improve imaging of brain functions. For example, in a recent collaboration with the Unit of Functional Imaging, we demonstrated for the first time the ability to measure neuronal activity directly with MRI techniques. These experiments pave the way to overcome current limitations of the MRI technique, which relies on measuring neuronal activity indirectly through oxygen consumption. We also have an ongoing collaboration with Dr. Pajevic (DCB/MSCL/OC) in which we develop new mathematical tools to analyze activity in large neuronal networks such as the cortex.